- Home
- Skills
- Commontoolsinc
- Labs
- Knowledge Base
knowledge-base_skill
- TypeScript
25
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill commontoolsinc/labs --skill knowledge-base- SKILL.md1.8 KB
Overview
This skill is a shared knowledge-base foundation for Oracle and Corrector agents that defines the authoritative source hierarchy for resolving conflicts between documentation, code, and specs. Load this skill first when investigating how the system works so agents apply a consistent priority order and surface unresolved contradictions.
How this skill works
The skill inspects available specifications, tests/patterns, runtime code, and plain-text docs to determine which source should be trusted for a given question. It enforces the rule that concrete artifacts outrank abstract guidance and provides a step-by-step decision flow so agents can justify their conclusions. When ambiguity remains, the skill marks the conflict for human review.
When to use it
- Investigating system behavior when documentation, tests, or code disagree
- Determining which source to follow when implementing or refactoring features
- Triaging bug reports that reference contradicting guides or specs
- Reviewing design decisions where intent and implementation diverge
- Preparing authoritative responses from Oracle or validating Corrector fixes
Best practices
- Always check docs/specs first for unambiguous intent before consulting other sources
- Use tests and pattern examples to confirm how the system actually behaves
- Treat runtime code as the source of truth for current behavior and edge cases
- Use plain-text docs only as learning material and validate any operational claims
- When conflict remains, explicitly surface the discrepancy and recommend human verification
Example use cases
- A user reports behavior that differs from a tutorial: verify tests and runtime code before updating the tutorial
- A spec describes a new API but runtime code hasn’t been updated: follow the spec for design discussions and note the implementation gap
- A pattern example demonstrates a workaround not present in the spec: prefer the test-proven approach for immediate fixes
- Automated agent producing a remediation: base changes on tests and runtime code, then adjust docs to match
FAQ
Fallback to working code and tests for concrete behavior. Flag the missing spec as an action item.
Which has higher authority: tests or runtime code?
Tests demonstrate intended behavior; runtime code shows actual behavior. Use tests to interpret intent, but trust runtime code for what the system currently does.